Basics Qualitative Research Grounded Theory Corbin Strauss

1. 1Read a segment of raw data (spoken words, actions, problems, issues) and name the analyst's impressionistic understanding of what is being described — this label is your concept.
2. 2Interrogate the concept for its properties (characteristics that define it) and dimensions (the range along which those properties vary), so the concept has substance rather than being a bare label.
3. 3Ask of diverse instances what they share in common — as with bird, plane, and kite sharing 'flight' — to find the abstracting principle that groups them.
4. 4Use that common concept to pull disparate data together, then explore each grouped instance in depth to detail similarities AND differences in terms of the concept.
5. 5Write a memo the moment a concept surfaces; keep early memos simple and focused on the single concept, then let later memos become summarylike, abstract, and integrative as you explore relationships between two or more concepts.
6. 6Sort memos by category; when a memo begins linking two or more concepts, treat that as a clue to integration rather than forcing it into one category.
7. 7Elevate only concepts that reach the status of major categories into diagrams, and keep balance — pursue the ideas worth developing, and let go before analysis collapses into boring minutia.

Corbin's own admission on the Vietnam veterans mini-project shows conceptualization done and corrected. As she reviewed her memos, she noticed some recorded her emotional response to the data rather than a conceptualization of what her respondents were telling her. That is the abstracting move failing: the memo captured her feeling, not the concept in the participant's account. She rewrote those memos so they abstracted the participant's view instead of her reaction — while still granting she could 'see myself in the analysis.' The tell was the memo content: reflective-of-self versus grouping-of-participant-meaning. This is conceptualization as a researcher-driven thinking and feeling process, using self-reflection to sharpen rather than slant the abstraction.

• If the goal is description, then do quality description and say so; if the goal is theory, then integrate concepts into an overarching explanatory scheme — do not do description and call it theory.
• If a concept only reaches minor status, then leave it out of integrative diagrams; only major categories belong there.
• If a memo reflects your emotional response more than what respondents are telling you, then rewrite it as a conceptualization of their view.
• If several diverse instances share an abstracting principle (their 'flight'), then group them under one concept, then explore each for similarities and differences.
• If reading the report becomes boring from over-detail, then you have overdeveloped — pull back toward conceptualization and away from minutia.

• Oversimplifying complexity out of existence — abstraction inevitably simplifies, and the tell is that the interpretation no longer reflects the bafflingly complex phenomenon it came from.
• Writing memos that record your feelings rather than the participants' meaning — the tell is memo content that is 'more reflective of my emotional response to the data than a conceptualization of what my respondents were telling me.'
• Producing thick description and labeling it theory — the tell is a report that leaves the reader confused about what is theory and what is not.
• Letting the computer program drive the analytic scheme — the tell is analysis restricted to the limits of a program's capabilities instead of the evolving analysis determining tool use.
• Muddled integrative diagrams crammed with too many words, lines, and arrows — the tell is a reviewer who cannot see the major point without a lot of explanation.

• ▸Data excerpt: paste the raw spoken words, action, or issue being analyzed
• ▸Concept name: your impressionistic label for what is being described
• ▸Common principle: what this and other instances share (the 'flight' test)
• ▸Properties: the defining characteristics of the concept
• ▸Dimensions: the range along which each property varies
• ▸Grouped instances: other data cases this concept pulls together, noting similarities and differences
• ▸Category status: minor / major (only major enters integrative diagrams)
• ▸Memo note: date-stamped analytic thought — flag if it links two or more concepts
• ▸Self-check: is this a conceptualization of the participant's view or my emotional response?

1. 1Start from a concept you have already identified in the data, then ask two separate questions of it: under what structural conditions does this appear (context), and what flow of action/interaction/emotion follows over time (process).
2. 2For context, locate the experience within the larger conditional frame in which it is embedded — specify the conditions (structural circumstances, situations, events) that shape the phenomenon rather than treating the experience as free-floating.
3. 3For process, trace the ongoing and changing forms of action/interaction/emotion that persons take in response to events and to the problems that arise to inhibit action/interaction.
4. 4Follow consequences forward: treat each consequence as feeding back into the next sequence of action, so process is analyzed as sequential and looping, not as a single static snapshot.
5. 5Use MAXQDA-style activation to hold context and process together: activate the relevant texts and a code with its subcodes (e.g., 'Survival' and its subcodes) so all coded segments across cases surface for comparison of conditions and sequences.
6. 6Write memos and draw diagrams as you go to capture how conditions link to action and to consequences, and revise the linkages as the analysis evolves — let the evolving analysis drive the tool, not the reverse.
7. 7Keep the balance Strauss warned about: abstract enough to order the complexity, but not so much that you distort the phenomenon by oversimplifying it out of existence.

Take the concept 'Survival' from the Vietnam veterans' study, coded with its subcodes across Participant #1 and Participants #2 in MAXQDA. Context analysis asks what conditions frame survival: combat situations, the structural realities of war, the surrounding events in which a veteran's actions are embedded. Process analysis then traces how the veteran's action/interaction/emotion changes over time — the responses to threatening events and to the problems that arise to inhibit action. Because consequences of one sequence become conditions for the next, you follow survival forward across the interviews and the written memoirs rather than freezing it at one moment. Activating the texts and the 'Survival' code cluster lets you retrieve every relevant segment and compare how conditions and sequences differ across participants. Memos attached in the margin record the emerging links between condition and action, which you diagram and revise as understanding deepens.

• If an explanation of experience leaves out the larger conditional frame, then it is incomplete — go back and specify context before proceeding.
• If you are describing what people do but not the changing form of that action over time, then you are missing process — add the action/interaction/emotion sequence.
• If a consequence appears, then treat it as a condition for the next sequence rather than an endpoint.
• If your abstraction is erasing the complexity Strauss valued, then you have oversimplified — pull back toward the data and thicker description.
• If your goal is theory, then integrate context and process into one explanatory scheme; if the goal is only description or theme-pulling, then say so and stop short of that integration.

• Treating context and process as one lumped analysis — the tell: you cannot say separately what conditions shape the phenomenon versus how action unfolds over time.
• Analyzing experience in isolation from its conditional frame — the tell: the account reads as free-floating meaning with no structural circumstances attached.
• Freezing process into a static snapshot — the tell: no sequence, no change over time, no feedback of consequences into later action.
• Letting the computer program direct the analysis — the tell: you restrict your thinking to what the program's capabilities allow instead of letting the evolving analysis decide how to use the tool.
• Calling integrated description 'theory' or unintegrated fragments 'analysis' — the tell: the reader is left confused about what is theory and what is not.

• ▸Concept: name the coded phenomenon you are analyzing (e.g., 'Survival').
• ▸Conditions/Context: list the structural circumstances and events that shape this phenomenon.
• ▸Conditional frame: state the larger context in which the experience is embedded.
• ▸Action/Interaction/Emotion: describe what persons do and feel in response to the event.
• ▸Change over time: note how that action/interaction/emotion shifts across the sequence.
• ▸Problems that inhibit action: identify what arises to block or complicate the action.
• ▸Consequences: record outcomes and how each feeds into the next sequence.
• ▸Memo link: reference the memo capturing the condition-action-consequence connection.
• ▸Goal check: mark whether this feeds description, concept analysis, or integrated theory.

1. 1Decide before you begin whether your goal is theory or description; if theory, commit to producing an overarching explanatory scheme rather than stopping at themes.
2. 2Identify a central or core category — the concept that appears most frequently, connects to the most other categories, and carries analytic power to account for variation in the data.
3. 3Link each major category to the core category, articulating how it conditions, extends, or is a consequence of the core phenomenon.
4. 4Use the paradigm as your integrating device: relate structure (context, conditions) to process (the flow of action/interaction/emotions in response to events).
5. 5Bring context and process together explicitly — trace how sets of conditions give rise to problems, how persons respond through action/interaction/emotion, and how those responses feed back to impact conditions.
6. 6Write memos and draw diagrams to display the relationships; move materials around (by hand or in a program like MAXQDA) until the categories cohere.
7. 7Refine the resulting theoretical formulation — test it against the data, fill gaps through theoretical sampling, and densify until it is coherent and explanatory.
8. 8Check the final scheme: if a reader cannot tell what is theory versus description, keep integrating rather than declaring the work finished.

In the Vietnam War study, Corbin takes the reader across Chapters 8–12 through the full sequence that culminates in integration. Concepts are first identified from veterans' interviews and written memoirs (Chapter 8), then elaborated (Chapter 9), contextualized (Chapter 10), and analyzed for process (Chapter 11). Only in Chapter 12 are the categories linked around a core category to build theory and development. The parallel biographical study of chest pain shows the logic in miniature: a life-threatening event (the condition) sets up conditions to which the person responds through action/interaction/emotion, and integration would relate those categories to how persons incorporate such events into their lives. The point is that integration is not a separate stroke of insight but the final analytic step resting on everything built before it. Corbin insists she is 'actually doing analysis' through these chapters rather than only describing it.

• If your goal is thick description, concept analysis, or theme-pulling, then stop before integration and say so — do not call the result theory.
• If several categories keep pointing back to one recurring concept, then treat that concept as your candidate core category.
• If you can relate structure to process, then use the paradigm as the analytic strategy that integrates the two.
• If the emerging scheme has gaps or thin categories, then return to theoretical sampling before attempting to finalize integration.
• If a reader would be confused about what is theory versus description, then the formulation is not yet integrated — keep refining.

• Doing description and calling it theory — the tell is that findings are listed as themes with no core category connecting them into an explanatory scheme.
• Stopping at concept identification or elaboration and treating that as the endpoint — the tell is categories that stand alone, unrelated to any central phenomenon.
• Letting the computer program dictate the analytic scheme — the tell is that analysis is restricted to the program's capabilities rather than the evolving analysis driving the tool's use.
• Committing to an analytic scheme too early — the tell is that categories are locked in before the data has been moved around and reconsidered.
• Treating integration as a mechanical recipe — the tell is a forced rather than skillful use of the tools, because the researcher has not thought and felt through the process.

• ▸Study goal: state 'theory' or 'description' — determines whether integration applies.
• ▸Candidate core category: the concept that recurs most and connects the most other categories.
• ▸Major categories: list each category to be related to the core.
• ▸Structural conditions (context): the sets of conditions in which the problem/situation arises.
• ▸Process: the flow of action/interaction/emotions persons use in response.
• ▸Consequences and feedback: how responses impact back on conditions.
• ▸Paradigm links: for each major category, how it conditions, extends, or results from the core.
• ▸Refinement notes: gaps to fill via theoretical sampling and memos/diagrams supporting the scheme.
• ▸Coherence check: can a reader distinguish theory from description here? yes/no + what to fix.

1. 1Take each main category and list its properties (its general characteristics) and dimensions (the range each property varies along). Ask of the category: what are its salient properties, and what is the range each can take?
2. 2Locate poorly developed categories by making diagrams and sorting memos — the gaps and thin spots show up when you try to represent relationships visually or arrange your memos into a story.
3. 3Fill in gaps first by reviewing existing memos and raw data for material you overlooked; a category may already be denser in your data than in your write-up.
4. 4When existing data won't fill the gap, return to the field and gather more through theoretical sampling — select sites, persons, events, or documents that will demonstrate the dimensional range or variation of that concept.
5. 5Build variation deliberately by comparing incidents and events on the basis of concepts (not description); since no two events are identical, differences in conditions, inter/action, or consequences give you dimensional range.
6. 6Locate experience within its larger framework — social, political, cultural, racial, gender-related, informational, technological — and treat process as integral, because action/interaction is formed and transformed by consequence and contingency.
7. 7Continue filling in through the writing phase, but stop when the category reaches theoretical saturation; decide what cannot be fully developed and let it go rather than chase every detail.

In Corbin's Vietnam War combatants study, the concept 'Homecoming' was recognized as important but was never developed to density — because of time constraints, Corbin did not gather the data needed to identify its properties and build in variation. The tell of this underdevelopment is exactly what the passages describe: a category that would appear thin when diagramming and sorting memos, with few specified conditions or dimensions attached to it. Contrast this with Corbin's work on women's perceived risks, where variation was built in by capturing that women defined risk differently from health professionals and acted on their own definitions. That variation — the same concept (risk) manifesting differently across actors and conditions — is what gave the analysis explanatory power. 'Homecoming' lacked that range of manifestations; 'risk' had it. The difference is not abstraction but density: how fully the properties and dimensions were specified across conditions.

• If a category looks thin when you diagram it or sort its memos, then treat it as poorly developed and fill it in before writing conclusions about it.
• If review of existing memos and raw data doesn't close the gap, then return to the field and sample theoretically for that specific category.
• If you cannot choose an ideal site or person, then make the most of available data and compare incidents on concepts anyway — theoretical sampling is defined by comparing on concepts, not by how data were gathered.
• If your analytic scheme 'doesn't quite seem right,' then suspect your own professional bias is overriding what respondents perceive, and return to the data with the respondents' definitions foreground.
• If time and resources force a choice, then accept that not every concept can reach density and stop at theoretical saturation rather than delaying indefinitely.

• Oversimplifying complexity out of existence — reducing multi-factor events to one explanation or scheme; the tell is a tidy analysis with no built-in variation and no unanticipated interactions.
• Leaving a recognized concept underdeveloped, like 'Homecoming' — the tell is a category you know matters but whose properties and dimensions are never specified.
• Being blinded by the researcher's own perspective — imposing your definitions (e.g., of risk) instead of respondents'; the tell is that the analytic scheme feels off and actions appear 'inconsistent' until you realign to respondents' meanings.
• Analyzing data descriptively rather than comparing incidents on concepts — the tell is that variation and process never emerge because you are cataloguing rather than dimensionalizing.
• Confusing volume of data with density — assuming more interviews equals saturation; the tell is broad coverage but shallow properties, gaps that only surface once you start writing.

• ▸Category name: the concept being examined
• ▸Properties: list the general characteristics of this category
• ▸Dimensions: for each property, state the range it varies along
• ▸Conditions specified: under what conditions does this concept appear/change?
• ▸Manifestations/variation: how does it differ across actors, sites, or events (name a real comparison)?
• ▸Larger framework located: which social/political/cultural/gender/technological aspects are attached?
• ▸Development source needed: memo review / raw data review / field theoretical sampling
• ▸Saturation status: saturated, needs filling, or accept-as-underdeveloped (with reason)

1. 1Decide your aim before you start: description, theme delineation, or theory. If theory, commit to integrating findings into an overarching explanatory scheme rather than stopping at description labeled as theory.
2. 2Build the conceptual pyramid from the bottom up. Identify lower-level concepts first, keeping them tied directly to the data, then group events under more abstract labels only after immersion yields insight.
3. 3Move up the conceptual ladder deliberately, understanding the trade-off: higher-level concepts gain explanatory breadth but lose specificity. Retain lower-level concepts in every explanation so you never drift too far from the data.
4. 4Locate a single overarching explanatory concept that stands above the rest — the one that, taken with the others, accounts for the what, how, when, where, and why of the phenomenon.
5. 5Test the explanation against the criteria of both art and science: it must construct a coherent story that 'feels right' yet not be a wild foray into fancy — you make, but cannot fake.
6. 6Situate experience within its conditional frame or context, and describe the process (ongoing, changing forms of action/interaction), since any explanation of experience is incomplete without both.
7. 7Preserve descriptive detail underneath the theory so the higher-level concepts rest on a solid foundation that goes back to the data, carrying the power of description with them.

Corbin illustrates explanatory power with the concept of flight applied to a bird, a plane, and a kite. At the abstract level, 'flight' is integral to each object — a kite that will not fly is a dud, a plane that will not fly is 'a car with wings.' But to actually understand and describe flight, you must examine the individual properties and dimensions of flight as they apply to each object, developing an understanding of flight under various conditions. The higher-level concept (flight) gains explanatory reach, yet only holds up because it rests on the lower-level, data-grounded details of each case. If you strip away those lower-level concepts, you keep the broad label but lose the specificity and the descriptive power that make the explanation compelling. This is why Corbin insists the conceptual pyramid be carefully crafted: the abstract explanation stays credible precisely because it can be traced straight back to the data.

• If your goal is description, then commit to 'quality' description and do not dress it up as theory — leaving the reader confused about what is and is not theory is a failure.
• If you aim to develop theory, then your findings must be systematically interrelated through statements of relationship and unified by one overarching explanatory concept.
• If a higher-level concept cannot be traced back down through lower-level concepts to the data, then it is too abstract to be trusted and needs regrounding.
• If more than one coherent story can be built from the data, then acknowledge the interpretation is a construction — but do not treat that as negating the relevance of the findings.
• If the story 'feels right' but you cannot show the analytic logic behind it, then you have art without science and cannot yet call it research.

• Doing description and calling it theory — the tell is the absence of any overarching explanatory concept interrelating the categories; findings sit as a list rather than a framework.
• Climbing to abstraction too fast — the tell is that concepts have lost specificity and no longer bring back detail or variation from the data.
• Early foreclosure: jumping to conclusions and grouping events under a label before immersion — the tell is conclusions that prove wrong as analysis proceeds.
• Becoming obsessed with coding procedures so the fluid, insight-driven nature of analysis is lost — the tell is a mechanically structured output that 'feels' dead and un-grounded.
• Faking coherence: making wild forays into fancy not supported by the data — the tell is a story that feels right but cannot survive being retraced through an audit trail back to the evidence.

• ▸Research aim: state whether this is description, theme delineation, or theory
• ▸Phenomenon under investigation: name the single thing being explained
• ▸Overarching explanatory concept: the one concept that stands above the rest
• ▸Supporting categories: list the well-developed categories that interrelate with it
• ▸Statements of relationship: how the categories connect (the what/how/when/where/why)
• ▸Lower-level grounding: for each concept, cite the data detail it rests on
• ▸Context/conditional frame: the larger conditions the experience is embedded in
• ▸Process description: the ongoing, changing forms of action/interaction
• ▸Coherence check: does it 'feel right' AND can you retrace it to the data (yes/no)?

1. 1Decide at the outset whether your goal is description or theory, and hold to it — do not present description and call it theory. Credibility claims must match the actual output.
2. 2Build credibility during the study, not after: take strategic action (theoretical sampling, sampling adequacy, active analytic stance, saturation) rather than waiting for external reviewers to proclaim it.
3. 3Provide sufficient detail and 'rich thick description' so readers feel they were vicariously in the field and can judge for themselves.
4. 4Make the process transparent: document how data were gathered and how analysis was conducted, and specify the kinds of data on which each interpretation rests, so readers can assess how you reached your conclusions.
5. 5Use multiple comparison groups where possible — Glaser and Strauss hold that findings based on more than one group have greater credibility.
6. 6Ground every concept in its context: locate expressed emotions, actions, and experiences within the conditions in which they occurred so meaning is clear and accurate; do not embellish participants' words to make a point.
7. 7Test applicability against four criteria: does the theory 'fit' the area it came from, is it understandable by laymen and professionals, is it general enough for diverse situations, and does it give the user enough control to bring about change.
8. 8Run Charmaz's self-evaluation across credibility, originality, resonance, and usefulness — but recognize self-evaluation is biased, so pair it with peer review or debriefing and member checks.

Strauss (1978) wanted a formal theory of 'negotiations.' Rather than confine data to one group, he examined negotiations across many contexts — representatives of nations, judges in law courts, political machines, clans and ethnic groups, and insurance companies with their clients. By comparing and contrasting these groups for similarities and differences, he identified components of the negotiation process that transcended all groups (raising abstraction and applicability) while also identifying the variations particular to each group. This use of multiple comparison groups is exactly what makes credibility greater, because findings rest on more than one group. It also delivers applicability: the resulting theory 'fits' each area, is general enough for diverse situations, and stays grounded in data even as it becomes more abstract. Contrast this with a substantive-level study confining data to one group — say, negotiations between buyer and seller in a housing transaction — which yields narrower applicability.

• If your goal is theory, then present findings as a set of interrelated concepts with an overarching explanatory scheme — not a list of themes; a listing is only description.
• If findings are both credible (believable, plausible) and applicable (they fit and are useful for diverse populations to bring about change), then the philosophic debate about 'truth,' 'validity,' and 'reliability' becomes superfluous — the proof is in the pudding.
• If you want broader applicability, then compare across multiple related contexts to raise the concept's abstraction; if you want depth in one group, then confine data collection to one main group and accept narrower reach.
• If you rely on self-evaluation criteria (Charmaz), then supplement with peer debriefing or member checks, because self-evaluation requires sophistication and still carries hard-to-remove bias.
• If you use a computer program, then use it to preserve a retraceable audit trail for reliability — but treat it as a tool, not a substitute for the thinking and feeling the analysis demands.

• Doing description and calling it theory — the tell: findings are a list of themes with no unifying explanatory scheme or relational statements linking concepts.
• Chasing only 'scientific' rigor criteria (triangulation, saturation, audit trails) while ignoring the creative aspects — the tell: no attention to context, process, density, variation, usefulness, or vividness and resonance.
• Taking words or actions out of context like a novelist to make a point — the tell: interpretations that contradict the data are smoothed over instead of presented fairly; meaning is distorted.
• Proclaiming credibility at project completion via external reviewers rather than attaining it through strategic action during the study — the tell: no evidence of theoretical sampling, active analytic stance, or saturation while data were being collected.
• Trusting self-evaluation alone — the tell: high researcher conviction (Glaser and Strauss note it becomes hard to shake) mistaken for proof, with no acknowledgment that another plausible analysis could rest on the same data.

• ▸Stated goal: description or theory? (name it in one line and confirm output matches)
• ▸Field-immersion evidence: what rich thick detail lets readers feel vicariously present?
• ▸Process transparency: how were data gathered and analysis conducted (audit trail entry)?
• ▸Data basis: which kinds of data does each key interpretation rest on?
• ▸Comparison groups: list every group compared; note what transcends them and what varies
• ▸Context check: is each emotion/action/experience located in the conditions where it occurred?
• ▸Fit: does the finding fit the area it came from and where it will be used?
• ▸Understandable & general: is it clear to laymen and professionals, and applicable to diverse situations?
• ▸Usefulness & control: does the analysis offer interpretations people can use to bring about change?

1. 1State your research aim at the outset and commit to it: description, thematic delineation, concept analysis, or theory building. This decision governs every downstream choice — do not blur the categories.
2. 2Ground your analysis in concepts. During data collection and analysis, examine and interpret data to elicit meaning; treat concept identification as the foundation, not an afterthought.
3. 3If your aim is theory, run the full sequence: concept identification, then concept elaboration, then contextualizing data (locating experience within its larger conditional frame), then bringing process (ongoing/changing action-interaction-emotion) into the analysis, then integration into an overarching theoretical explanatory scheme.
4. 4Look for consequences at each analytic pass, because consequences feed back into the next sequence of action — build this loop into your coding rather than stopping at description.
5. 5Test provisional findings against peers. Following the Pragmatist model, treat new knowledge as provisional until checked out empirically; validity lies in the consequences of the operations you perform.
6. 6Use tools — comparisons, generative questions, memos, diagrams, computer programs — to support the thinking, but keep analysis researcher-driven. Let the evolving analysis determine how you use a program, never the reverse.
7. 7Write findings so the reader knows exactly what they are receiving: if it is description, deliver quality description; if it is theory, deliver an integrated explanatory scheme. Do not label description as theory.

In the 3rd edition of Basics, Corbin demonstrates 'the doing of analysis' using materials from the Vietnam War as the research project, and the chapter sequence itself models how empirical knowledge accumulates. Chapter 8 begins with concept identification, Chapter 9 moves to concept elaboration, Chapter 10 contextualizes the data, Chapter 11 brings process into the analysis, and Chapter 12 integrates the pieces into theory. Each step is not a summary but an actual worked analytic operation on the same body of data. The payoff is twofold as she states: the concepts increase understanding of persons in their everyday lives — their routines, habits, problems — and they supply a shared language for discussion and debate. That shared language is what lets a discipline build a professional body of knowledge and enhance practice. The Vietnam materials thus serve as a concrete demonstration that concepts derived from data are the currency of cumulative, evidence-based understanding.

• If your goal is theory, then integrate findings into an overarching explanatory scheme; if you stop at concepts or themes, then say so explicitly and do not call it theory.
• If you are tempted to oversimplify to make abstraction clean, then check Strauss's balance question — preserve enough complexity that interpretation is not distorted out of existence.
• If a computer program's limits are shaping your coding, then override the program — restrict analysis to your evolving thinking, not to the software's capabilities.
• If a finding feels final, then treat it as provisional and submit it to peers for empirical checking before claiming it as knowledge.
• If explaining an experience, then also locate it in its conditional context and describe the process of action/interaction/emotion — an explanation lacking both is incomplete.

• Doing description and calling it theory. The tell: the reader cannot distinguish what is theory from what is not, and there is no integrated explanatory scheme.
• Collapsing into navel-gazing and 'telling nice stories' at the expense of building a professional body of empirical knowledge. The tell: the purpose of the research — contributing verifiable understanding — has quietly disappeared.
• Treating the book (or any method) as a recipe. The tell: rigid, step-locked procedure with no researcher judgment; Corbin says she'd be offended by this reading.
• Letting the computer program direct the process. The tell: analysis is truncated to what the software supports, and the researcher-driven thinking-and-feeling process is rigidified.
• Skipping the philosophy-of-knowledge grounding, so methods float free of any epistemology. The tell: you cannot say why methodologically you do what you do, only that you follow steps.

• ▸Research aim: state one — description / themes / concept analysis / theory
• ▸Philosophical orientation: name the worldview informing your methods (e.g., Interactionism/Pragmatism)
• ▸Core concepts identified: list the concepts your analysis rests on
• ▸Context frame: describe the larger conditions in which the experience is embedded
• ▸Process traced: describe the changing action/interaction/emotion in response to events
• ▸Consequences logged: note how consequences feed the next sequence of action
• ▸Integration status: is this integrated into an explanatory scheme, or left as concepts/themes?
• ▸Peer check plan: how and by whom will findings be checked empirically
• ▸Contribution statement: one line on what shared understanding this adds to the field's body of knowledge

1. 1Write a philosophical self-location statement before you code any data: name where you sit among post-positivist, constructivist, and pragmatist positions, as Corbin does when she recounts being labeled a 'post-positivist' and then absorbing valid points from the postmodern schools.
2. 2Trace your training lineage explicitly. Corbin names herself 'trained as a grounded theorist' under Strauss, and distinguishes procedures inherited from Glaser/Strauss (comparisons, generative questions, theoretical sampling) that any orientation can use from theory-laden moves that require adopting the Pragmatist/Interactionist stance.
3. 3Adopt the Pragmatist ontology of a fluid, indeterminate world where 'organism and environment act together, or inter-act' (Dewey, 1938): design to capture action, interaction, and emotions as ongoing changing forms, plus consequences that feed the next sequence.
4. 4Declare your research aim up front — description, concept analysis, theme-pulling, case study, change, or theory building — because the passage insists you 'be very clear at the beginning of a study what it is they are setting out to do' and never call description 'theory.'
5. 5Practice reflexivity as an operational step, not an attitude: keep a methodological log of decisions and write reflective memos, then flag which memos capture your emotional response versus conceptualization of respondents (Corbin's own tell during the Vietnam veterans study).
6. 6Treat your personal traits as instruments: identify the curiosity, tolerance for ambiguity, and willingness to let respondents 'talk for themselves' that the qualitative revolution requires, and name how each will shape your fieldwork.
7. 7Position technology as support, not driver: if using a computer program, state that it 'supports' rather than 'takes over' or 'directs' analysis, keeping the process 'researcher-driven thinking and feeling.'

Corbin models this construct through her own confession. Trained in a single 'grounded theory' approach under Strauss, she one day found herself labeled a 'post-positivist' and watched objectivity dismissed and reality-in-data called a fantasy during the Qualitative Revolution. Rather than defending her old bubble, she interrogated her aptitude and worldview: she affirmed her Pragmatist and Interactionist orientation and her nursing-born practical aim to 'develop knowledge that will guide practice,' plus a feminist social-justice aim. She then reconciled Strauss's vision of theory-building with her own belief that thick description, case analysis, change, and storytelling are equally valid. In her Vietnam veterans mini-study she found self-reflection 'cathartic' and used it to catch herself 'slanting the data.' The orientation is not decoration — it is what let her keep integrity while telling participants' stories.

• If your goal is theory, then integrate findings into 'an overarching theoretical explanatory scheme'; if it is description, then do 'quality' description and say so — never blur the two.
• If a procedure came from Glaser/Strauss (comparisons, generative questions, theoretical sampling), then use it regardless of orientation; if a move depends on the Pragmatist/Interactionist view of a fluid, action-driven world, then only claim it after adopting that stance.
• If you cannot separate who you are from the research, then treat reflexivity as mandatory and log position, perspective, and presence — but decide yourself 'how much, when, and how' since feasibility is debated.
• If a memo reads as your emotional response rather than a conceptualization of the respondent, then mark it as reflexive data, not analytic finding.
• If a computer program starts rigidifying your coding, then override it: let the evolving analysis determine the tool's use, not the reverse.

• Doing description and calling it theory — the tell is a reader left 'confused about what is theory and what is not' because findings were never integrated into an explanatory scheme.
• Treating the method as a recipe — the tell is mechanical procedure-following; the author says she would be 'offended' if the book is read as a recipe because analysis is something you 'feel yourself through.'
• Navel-gazing reflexivity that displaces the study — Corbin's fear that researchers become 'so concerned with examining their own navels and telling nice stories' that they lose the purpose of generating empirical knowledge.
• Applying theory dogmatically — the ICU-nurse test: a practitioner who applies principles without reevaluating and adjusting 'to meet the situation at hand' has mistaken orientation for a rulebook.
• Letting the computer program direct analysis — the tell is a rigidified process; the passage warns the tool should 'support,' not 'take over' the researcher-driven thinking and feeling.

• ▸Training lineage: name the methodology and mentor you were trained in.
• ▸Current philosophical position: post-positivist / constructivist / Pragmatist-Interactionist / other — and why.
• ▸Ontological assumption: one sentence on the nature of the world you study (e.g., fluid, action-driven, multiple perspectives).
• ▸Research aim declared: description / concept analysis / themes / case study / change / theory building.
• ▸Practical purpose: what change or knowledge-for-practice you want the study to serve.
• ▸Personal traits as instrument: list curiosity, tolerance for ambiguity, creativity you bring.
• ▸Reflexivity plan: how much, when, and how you will log decisions and separate emotion from conceptualization.
• ▸Procedure vs. stance check: which techniques you use are orientation-neutral vs. orientation-dependent.
• ▸Technology stance: one line stating the computer program supports rather than directs your analysis.

1. 1Train and rehearse before entering the field: interviewing and observing are skills that require practice, not tasks you can improvise on arrival, or you will return with sparse data.
2. 2Default to the unstructured interview when you want density: open with an invitation like 'Tell me about your experience with X in your own words' and hold questions until after the participant's uninterrupted story.
3. 3Sit with an open mind and open agenda; do not let nervousness push you into filling silence or steering, because that blocks the free flow of information.
4. 4Triangulate sources deliberately: combine interviews with observation, then add documents, memoirs, diaries, or reports to verify and add perspectives on the same problem.
5. 5Deliberately gather multiple perspectives on the topic and bring context into the account, since context helps define and create the situations and responses you are trying to explain.
6. 6Watch for language that carries dense meaning — metaphors participants use ('going through hell,' 'fighting a battle') — and record their exact words rather than paraphrasing them away.
7. 7In cross-cultural work, expect complexity to be lost in translation; preserve source-language concepts and use cultural advisors to help you read what participants mean.

Noriko Yamamoto, studying Japanese family caregivers of parents with dementia, captured two source-language concepts her data revealed that had no English equivalent: amaeru, the younger caregiver seeking indulgent love from a care recipient still able to respond, and amayakasu, the caregiver offering indulgent love once the recipient can no longer respond as an adult (Yamamoto & Wallhagen, 1998). Had she translated these into a rough English gloss, the distinction between seeking and offering indulgent love — and how it shifts as dementia increases — would have collapsed. The richness of her raw data came precisely from preserving the participants' own terms. Compare Chesney (2001), who used Pakistani advisors to understand what her participants were telling her and still wished she were Pakistani to fully bridge the barrier. Both cases show that quality of raw data depends on capturing meaning as participants frame it, not as the researcher's language allows.

• If you want the most data-dense material, then use an unstructured interview with no predetermined question set rather than a fixed schedule.
• If your participants are 'less than informative' and data are sparse, then the problem is likely lack of interviewer training and preparation, not the participants.
• If a concept in your data has no clean English equivalent, then keep the source-language term and define it, rather than forcing a translation that flattens complexity.
• If one source cannot verify a claim, then add another type of data (observation, documents) to triangulate the same problem.
• If you are working across cultures or genders, then assume some complexity will be lost and build in advisors or member checks to recover it.

• Assuming you can walk into the field and interview or observe with no training — the tell is disappointing sessions where participants say little and notes are thin.
• Over-structuring interviews with a predetermined question list, which suppresses the free flow of information and yields shallow answers.
• Letting nervousness fill silences and steer the participant, so you record your agenda instead of theirs.
• Translating away distinctive participant language or source-culture concepts, losing the meaning that made the datum rich — the tell is a transcript that sounds like the researcher, not the participant.
• Rushing the project to save time, energy, or money and narrowing to a single perspective, which strips out the depth, breadth, and contextual variation that quality depends on.

• ▸Source type: interview / observation / document / memoir / other — name which one.
• ▸Format used: unstructured / semi-structured / structured — note the actual opening prompt.
• ▸Interviewer preparation: what training or rehearsal preceded this collection?
• ▸Participant's own words captured: paste 1-2 verbatim phrases or metaphors that carry meaning.
• ▸Untranslatable/source-language concepts: list any terms kept in original language and why.
• ▸Context recorded: what situational factors (time, place, relationships) were noted?
• ▸Perspectives covered: which viewpoints on the topic does this source add?
• ▸Triangulation link: which other source verifies or extends this one?
• ▸Thinness flag: mark 'sparse' if answers were short or steered, and note the likely cause.

1. 1State your goal before touching data: decide whether you are producing thick description, concept analysis, theme extraction, or full theory construction. This choice governs how far you push the procedures — you can stop at concepts without integrating them into an 'overarching theoretical explanatory scheme,' as long as you say so.
2. 2Code the data to move from raw text to concepts. Break the material apart, name what is happening, and identify recurring ideas — for the Vietnam material this means reading a veteran's account and labeling processes like 'graduated pressure,' 'containment,' and 'sense of powerlessness.'
3. 3Compare incidents against incidents and concept against concept to find similarities and differences; this comparison is what separates description from conceptualization.
4. 4Question the data actively — ask what, how, why, and 'what if it were otherwise' — to keep interpretation from settling into an analytic scheme too early.
5. 5Write memos as you code to record your thinking, and draw diagrams to show relationships among concepts; keep both retrievable so the reasoning is transparent.
6. 6If your goal is theory, integrate the categories into a single explanatory scheme — do not stop at a list of themes and mislabel it theory.
7. 7Optionally load the data into a computer program (MAXQDA in this book) to store, sort, search, and retrieve codes and memos, and to move materials around so you can rethink them — but let the evolving analysis drive the program, never the reverse.
8. 8Build an audit trail by retracing your codes, memos, and diagrams so another reader can evaluate how you got from data to findings.

Corbin analyzes McNamara's conduct of the Vietnam War from source materials. Coding the passage yields concepts: 'graduated pressure' (slow, controlled bombing rather than major assault), 'containment' (a plan not aimed at winning), and 'civilian understanding of war' (business logic applied to combat). Comparison surfaces a relationship — every proposed action had to be relayed to Washington for approval, so military advantage was lost, and the Joint Chiefs were given little input. Questioning the data ('why did this fail?') exposes two flaws: the plan sought negotiation not victory, and it ignored that North Vietnam had 'little infrastructure to bomb.' A memo would record that these codes cluster around a category like 'powerlessness of operators under distant control,' and a diagram would link approval delay to lost advantage to that sense of powerlessness. Stopping here gives rich concept analysis; integrating these categories into one explanatory scheme would move it toward theory.

• If your goal is description, then do quality description and label it as such — do not integrate categories into a theoretical scheme and call it theory.
• If you keep finding new dimensions when comparing incidents, then keep sampling and coding; if comparison yields no new properties, then the category is saturating.
• If a computer program's structure is starting to dictate what you code, then stop and let the evolving analysis decide — the program supports, it does not direct.
• If you commit to an analytic scheme early and force later data into it, then loosen it; use the ability to move materials around to rethink relationships.
• If someone will need to judge reliability or follow an audit trail, then keep memos, codes, and diagrams retrievable and retraceable from the start.

• Doing description but presenting it as theory — the tell is a set of themes with no overarching explanatory scheme integrating them, leaving readers unsure what is theory.
• Treating the book as a recipe and applying procedures mechanically — the tell is rigid, lockstep coding with no thinking or feeling driving it.
• Letting the computer program take over: rigidifying analysis to fit the program's capabilities — the tell is analysis restricted to 'the limits of a program's capabilities.'
• Committing to an analytic scheme too early — the tell is that new data gets forced into existing codes instead of prompting new ways to think about the material.
• Skipping memos and diagrams, so the reasoning behind concepts is lost — the tell is no audit trail and an inability to retrace how findings emerged.

• ▸Study goal: state description, concept analysis, theme extraction, or theory
• ▸Data segment: the passage or incident being examined
• ▸Codes/concepts: names you assigned to what is happening
• ▸Comparison note: what this incident shares with / differs from others
• ▸Question asked: the what/how/why/what-if you posed to the data
• ▸Memo: your current thinking about meaning and relationships
• ▸Diagram reference: where the visual of concept relationships lives
• ▸Integration status: are categories being pulled into one explanatory scheme? yes/no
• ▸Audit note: how to retrace this step later (program location, file, date)

1. 1Begin data collection without a fixed sampling frame; let your first interviews or field notes generate concepts before you decide who or what to sample next. In grounded theory the analysis, not a preset design, drives collection.
2. 2After coding your initial data, identify the emerging concepts and the theoretical questions they raise (properties, dimensions, conditions, consequences that are still thin or unexplained), and write these into memos so the sampling decision is explicit and traceable.
3. 3Decide what data to collect next and where to find them based on those emerging concepts — go seek the incidents, people, or situations that will develop a category's properties and dimensions, not a representative cross-section.
4. 4Use the technical literature deliberately as a source for theoretical sampling: let a concept recalled from published work suggest a comparison case or a new site to sample (Ch. 4 lists 'suggest areas for theoretical sampling' as a use of literature).
5. 5Make comparisons at the property and dimensional level between the new data and prior data, asking generative questions about differences and similarities to sharpen each category.
6. 6Iterate: collect, code, memo, then re-decide the next sampling target based on the newly emerged concepts. Continue until categories are well developed and further sampling stops yielding new properties.
7. 7If your aim is a full theory, integrate the sampled concepts into an overarching explanatory scheme; if your aim is thick description or theme analysis, you can stop earlier but must say plainly you are not building theory.

Suppose you study 'coping' with the loss of a spouse due to a workplace accident. Your first interviews surface a concept of 'loss' with properties like timing of loss, previous experience with loss, and feelings about the spouse. Comparing this against the literature on other coping situations, you notice loss from divorce is conceptually similar but different. That emerging concept becomes a theoretical question — how does the suddenness and blamelessness of accidental death change coping? — which tells you where to sample next: you deliberately seek participants who lost a spouse through divorce so you can compare the two situations on the same properties and dimensions. The comparison delineates important features of each type of loss that a single sample would never reveal. Here the literature-derived concept and the emerging data jointly directed the next round of collection, which is theoretical sampling in action.

• If a category is still thin on properties or dimensions, then sample the next case specifically to develop that category — not to add generic volume.
• If a concept from the literature seems similar or opposite to one emerging in your data, then treat it as a suggested area for theoretical sampling and seek a comparison case, but compare only at the property/dimensional level, never as data itself.
• If new data stops producing new properties or dimensions for your categories, then that line of sampling is saturated — redirect to an underdeveloped category or move toward integration.
• If your goal is description or theme extraction rather than theory, then you may stop the iterative sampling earlier, provided you state clearly you are not building theory.
• If a computer program's structure is steering which data you pursue, then override it — the evolving analysis determines how you use the tool, not the reverse.

• Designing the full sample up front and collecting it before analysis — the tell: you cannot point to a memo showing which emerging concept prompted each new case; sampling was driven by design, not by concepts.
• Treating literature concepts as data rather than as comparison prompts — the tell: you import published findings directly into your categories instead of comparing them at the property and dimensional level.
• Sampling for volume or representativeness instead of for concept development — the tell: more interviews but no new properties or dimensions emerging.
• Calling the result 'theory' when it is description with unlinked themes — the tell: no overarching explanatory scheme integrating the sampled concepts, leaving readers unsure what is theory and what is not.
• Letting the software rigidify the process — the tell: analysis is restricted to the program's categories and you never move materials around or rethink them, so sampling decisions follow the tool rather than the emerging analysis.

• ▸Emerging concept: name the concept your current data raised.
• ▸Property/dimension gap: state what is still thin or unexplained about it.
• ▸Theoretical question: write the question this gap poses.
• ▸Literature prompt: note any published concept suggesting a comparison or site (optional).
• ▸Next data target: specify who/what/where you will sample and why it develops this concept.
• ▸Comparison plan: name the property/dimension you will compare across cases.
• ▸Saturation check: record whether the last round produced new properties (yes/no).
• ▸Aim tag: mark whether this feeds theory-building or description/themes.

1. 1Name your paradigm before coding: write down the perspectives, training, professional knowledge, and biases you bring, since these become woven into every stage. You are not emptying your head — Dey's rule is 'an open mind, not an empty head.'
2. 2Alternate data collection and analysis rather than collecting everything first. Sensitivity develops through this back-and-forth; meanings and significances emerge only as you cycle between people and data.
3. 3When you hit a datum you cannot interpret, treat existing knowledge and personal experience as a comparative base — ask what this resembles that you already understand — but always compare that knowledge back against the data, never losing sight of the data themselves.
4. 4Work every concept in terms of its properties and dimensions, so that a participant's meaning stands out when contrasted against other participants' descriptions or your own experience.
5. 5Reflect on your memos after writing them: separate memos that are your emotional response to the data from memos that conceptualize what respondents are telling you. Rewrite the emotional ones.
6. 6Test each interpretation against the 'Aha' threshold — you can say 'that is what they are telling me' only when the reading is anchored in data, not forced by your ideas talking through your own eyes.
7. 7Carry unresolved analytic problems with you off the desk; insight often arrives while reading the paper, talking to a colleague, or waking from a dream, then gets applied back to the data.

Corbin describes developing sensitivity while running the mini-project on Vietnam veterans. She found self-reflection natural and cathartic, and it let her see how she was slanting the data. On review, some of her memos turned out to be her own emotional reactions to the veterans' accounts rather than conceptualizations of what the veterans were telling her. She rewrote those memos so the analysis reflected the participants' views, while acknowledging she could still 'see herself in the analysis.' The passages also give the grief-and-loss case: a researcher who never experienced an unwanted divorce but who has lost a loved one uses that experience to build a list of general properties of loss and grief, then uses those properties as a comparative base to examine divorce data. In both cases experience supplies the starting insight, but only the data are finally significant.

• If a memo records mostly your emotional response rather than a conceptualization of what participants said, then rewrite it as conceptualization.
• If you feel it is 'only the data talking,' then check whether it is data talking through your eyes — the more you name your subjectivity, the less you force ideas onto data.
• If prior experience lets you grasp significance faster (skipping weeks of gaining familiarity), then use it; if a fresh outlook matters more for this site, then favor the new researcher's naivety.
• If knowledge and data conflict, then the data win — always compare experience against data and never lose sight of the data themselves.
• If a datum has no meaning yet, then hold it in mind across daily life and wait for the connecting insight rather than forcing a code.

• Claiming objectivity — treating findings as pure data talking. The tell: no account of your paradigm, perspective, or biases anywhere in the record.
• Forcing (Glaser's term): imposing your preconceived ideas on the data. The tell: categories arrive fully formed before immersion, and disconfirming data get ignored.
• Emptying your head instead of opening your mind — dispensing with accumulated knowledge. The tell: you re-derive obvious meanings from scratch and stall, unable to see connections.
• Letting experience blind you: prior familiarity makes you read data 'correctly' too fast and miss nuance. The tell: you skip properties and dimensions and assert meaning without comparison.
• Doing rich description and calling it theory. The tell: findings are never integrated into an overarching explanatory scheme, yet the write-up claims to build theory.

• ▸Paradigm brought in: name your perspectives, training, and biases for this study
• ▸Prior experience tapped: what known thing are you comparing this datum against
• ▸Datum under analysis: the specific quote or observation
• ▸Emotional response: your gut reaction (to be quarantined, not coded)
• ▸Conceptualization: what participants are actually telling you, in concept terms
• ▸Properties/dimensions: how this concept varies against other participants
• ▸Data check: how this reading is anchored in the data, not forced
• ▸Aha status: can you honestly say 'that is what they are telling me'? (Y/N)
• ▸Rewrite needed: does this memo mix emotion and conceptualization? (Y/N)